1 """Macintosh binhex compression/decompression.
4 binhex(inputfilename, outputfilename)
5 hexbin(inputfilename, outputfilename)
9 # Jack Jansen, CWI, August 1995.
11 # The module is supposed to be as compatible as possible. Especially the
12 # easy interface should work "as expected" on any platform.
13 # XXXX Note: currently, textfiles appear in mac-form on all platforms.
14 # We seem to lack a simple character-translate in python.
15 # (we should probably use ISO-Latin-1 on all but the mac platform).
16 # XXXX The simple routines are too simple: they expect to hold the complete
17 # files in-core. Should be fixed.
18 # XXXX It would be nice to handle AppleDouble format on unix
19 # (for servers serving macs).
20 # XXXX I don't understand what happens when you get 0x90 times the same byte on
21 # input. The resulting code (xx 90 90) would appear to be interpreted as an
22 # escaped *value* of 0x90. All coders I've seen appear to ignore this nicety...
29 __all__ = ["binhex","hexbin","Error"]
31 class Error(Exception):
34 # States (what have we written)
35 [_DID_HEADER, _DID_DATA, _DID_RSRC] = range(3)
38 REASONABLY_LARGE=32768 # Minimal amount we pass the rle-coder
40 RUNCHAR=chr(0x90) # run-length introducer
43 # This code is no longer byte-order dependent
46 # Workarounds for non-mac machines.
48 from Carbon.File import FSSpec, FInfo
49 from MacOS import openrf
51 def getfileinfo(name):
52 finfo = FSSpec(name).FSpGetFInfo()
53 dir, file = os.path.split(name)
54 # XXX Get resource/data sizes
58 fp = openrf(name, '*rb')
61 return file, finfo, dlen, rlen
63 def openrsrc(name, *mode):
68 return openrf(name, mode)
72 # Glue code for non-macintosh usage
81 def getfileinfo(name):
83 # Quick check for textfile
85 data = open(name).read(256)
87 if not c.isspace() and (c<' ' or ord(c) > 0x7f):
94 dir, file = os.path.split(name)
95 file = file.replace(':', '-', 1)
96 return file, finfo, dsize, 0
99 def __init__(self, *args):
102 def read(self, *args):
105 def write(self, *args):
111 class _Hqxcoderengine:
112 """Write data to the coder in 3-byte chunks"""
114 def __init__(self, ofp):
118 self.linelen = LINELEN-1
120 def write(self, data):
121 self.data = self.data + data
122 datalen = len(self.data)
123 todo = (datalen//3)*3
124 data = self.data[:todo]
125 self.data = self.data[todo:]
128 self.hqxdata = self.hqxdata + binascii.b2a_hqx(data)
131 def _flush(self, force):
133 while first <= len(self.hqxdata)-self.linelen:
134 last = first + self.linelen
135 self.ofp.write(self.hqxdata[first:last]+'\n')
136 self.linelen = LINELEN
138 self.hqxdata = self.hqxdata[first:]
140 self.ofp.write(self.hqxdata + ':\n')
145 self.hqxdata + binascii.b2a_hqx(self.data)
150 class _Rlecoderengine:
151 """Write data to the RLE-coder in suitably large chunks"""
153 def __init__(self, ofp):
157 def write(self, data):
158 self.data = self.data + data
159 if len(self.data) < REASONABLY_LARGE:
161 rledata = binascii.rlecode_hqx(self.data)
162 self.ofp.write(rledata)
167 rledata = binascii.rlecode_hqx(self.data)
168 self.ofp.write(rledata)
173 def __init__(self, (name, finfo, dlen, rlen), ofp):
174 if type(ofp) == type(''):
176 ofp = open(ofname, 'w')
179 fss.SetCreatorType('BnHq', 'TEXT')
180 ofp.write('(This file must be converted with BinHex 4.0)\n\n:')
181 hqxer = _Hqxcoderengine(ofp)
182 self.ofp = _Rlecoderengine(hqxer)
188 self._writeinfo(name, finfo)
189 self.state = _DID_HEADER
191 def _writeinfo(self, name, finfo):
194 raise Error, 'Filename too long'
195 d = chr(nl) + name + '\0'
196 d2 = finfo.Type + finfo.Creator
198 # Force all structs to be packed with big-endian
199 d3 = struct.pack('>h', finfo.Flags)
200 d4 = struct.pack('>ii', self.dlen, self.rlen)
201 info = d + d2 + d3 + d4
205 def _write(self, data):
206 self.crc = binascii.crc_hqx(data, self.crc)
210 # XXXX Should this be here??
211 # self.crc = binascii.crc_hqx('\0\0', self.crc)
216 self.ofp.write(struct.pack(fmt, self.crc))
219 def write(self, data):
220 if self.state != _DID_HEADER:
221 raise Error, 'Writing data at the wrong time'
222 self.dlen = self.dlen - len(data)
225 def close_data(self):
227 raise Error, 'Incorrect data size, diff=%r' % (self.rlen,)
229 self.state = _DID_DATA
231 def write_rsrc(self, data):
232 if self.state < _DID_DATA:
234 if self.state != _DID_DATA:
235 raise Error, 'Writing resource data at the wrong time'
236 self.rlen = self.rlen - len(data)
240 if self.state < _DID_DATA:
242 if self.state != _DID_DATA:
243 raise Error, 'Close at the wrong time'
246 "Incorrect resource-datasize, diff=%r" % (self.rlen,)
252 def binhex(inp, out):
253 """(infilename, outfilename) - Create binhex-encoded copy of a file"""
254 finfo = getfileinfo(inp)
255 ofp = BinHex(finfo, out)
257 ifp = open(inp, 'rb')
258 # XXXX Do textfile translation on non-mac systems
266 ifp = openrsrc(inp, 'rb')
274 class _Hqxdecoderengine:
275 """Read data via the decoder in 4-byte chunks"""
277 def __init__(self, ifp):
281 def read(self, totalwtd):
282 """Read at least wtd bytes (or until EOF)"""
286 # The loop here is convoluted, since we don't really now how
287 # much to decode: there may be newlines in the incoming data.
289 if self.eof: return decdata
291 data = self.ifp.read(wtd)
293 # Next problem: there may not be a complete number of
294 # bytes in what we pass to a2b. Solve by yet another
299 decdatacur, self.eof = \
300 binascii.a2b_hqx(data)
302 except binascii.Incomplete:
304 newdata = self.ifp.read(1)
307 'Premature EOF on binhex file'
308 data = data + newdata
309 decdata = decdata + decdatacur
310 wtd = totalwtd - len(decdata)
311 if not decdata and not self.eof:
312 raise Error, 'Premature EOF on binhex file'
318 class _Rledecoderengine:
319 """Read data via the RLE-coder"""
321 def __init__(self, ifp):
324 self.post_buffer = ''
328 if wtd > len(self.post_buffer):
329 self._fill(wtd-len(self.post_buffer))
330 rv = self.post_buffer[:wtd]
331 self.post_buffer = self.post_buffer[wtd:]
334 def _fill(self, wtd):
335 self.pre_buffer = self.pre_buffer + self.ifp.read(wtd+4)
337 self.post_buffer = self.post_buffer + \
338 binascii.rledecode_hqx(self.pre_buffer)
343 # Obfuscated code ahead. We have to take care that we don't
344 # end up with an orphaned RUNCHAR later on. So, we keep a couple
345 # of bytes in the buffer, depending on what the end of
346 # the buffer looks like:
347 # '\220\0\220' - Keep 3 bytes: repeated \220 (escaped as \220\0)
348 # '?\220' - Keep 2 bytes: repeated something-else
349 # '\220\0' - Escaped \220: Keep 2 bytes.
350 # '?\220?' - Complete repeat sequence: decode all
351 # otherwise: keep 1 byte.
353 mark = len(self.pre_buffer)
354 if self.pre_buffer[-3:] == RUNCHAR + '\0' + RUNCHAR:
356 elif self.pre_buffer[-1] == RUNCHAR:
358 elif self.pre_buffer[-2:] == RUNCHAR + '\0':
360 elif self.pre_buffer[-2] == RUNCHAR:
365 self.post_buffer = self.post_buffer + \
366 binascii.rledecode_hqx(self.pre_buffer[:mark])
367 self.pre_buffer = self.pre_buffer[mark:]
373 def __init__(self, ifp):
374 if type(ifp) == type(''):
377 # Find initial colon.
382 raise Error, "No binhex data found"
383 # Cater for \r\n terminated lines (which show up as \n\r, hence
384 # all lines start with \r)
390 dummy = ifp.readline()
392 hqxifp = _Hqxdecoderengine(ifp)
393 self.ifp = _Rledecoderengine(hqxifp)
397 def _read(self, len):
398 data = self.ifp.read(len)
399 self.crc = binascii.crc_hqx(data, self.crc)
403 filecrc = struct.unpack('>h', self.ifp.read(2))[0] & 0xffff
404 #self.crc = binascii.crc_hqx('\0\0', self.crc)
405 # XXXX Is this needed??
406 self.crc = self.crc & 0xffff
407 if filecrc != self.crc:
408 raise Error, 'CRC error, computed %x, read %x' \
412 def _readheader(self):
414 fname = self._read(ord(len))
415 rest = self._read(1+4+4+2+4+4)
420 flags = struct.unpack('>h', rest[9:11])[0]
421 self.dlen = struct.unpack('>l', rest[11:15])[0]
422 self.rlen = struct.unpack('>l', rest[15:19])[0]
426 self.FInfo.Creator = creator
427 self.FInfo.Type = type
428 self.FInfo.Flags = flags
430 self.state = _DID_HEADER
433 if self.state != _DID_HEADER:
434 raise Error, 'Read data at wrong time'
437 n = min(n, self.dlen)
442 rv = rv + self._read(n-len(rv))
443 self.dlen = self.dlen - n
446 def close_data(self):
447 if self.state != _DID_HEADER:
448 raise Error, 'close_data at wrong time'
450 dummy = self._read(self.dlen)
452 self.state = _DID_DATA
454 def read_rsrc(self, *n):
455 if self.state == _DID_HEADER:
457 if self.state != _DID_DATA:
458 raise Error, 'Read resource data at wrong time'
461 n = min(n, self.rlen)
464 self.rlen = self.rlen - n
469 dummy = self.read_rsrc(self.rlen)
471 self.state = _DID_RSRC
474 def hexbin(inp, out):
475 """(infilename, outfilename) - Decode binhexed file"""
482 out = ofss.as_pathname()
484 ofp = open(out, 'wb')
485 # XXXX Do translation on non-mac systems
493 d = ifp.read_rsrc(128000)
495 ofp = openrsrc(out, 'wb')
498 d = ifp.read_rsrc(128000)
504 nfinfo = ofss.GetFInfo()
505 nfinfo.Creator = finfo.Creator
506 nfinfo.Type = finfo.Type
507 nfinfo.Flags = finfo.Flags
508 ofss.SetFInfo(nfinfo)
514 binhex(fname, fname+'.hqx')
515 hexbin(fname+'.hqx', fname+'.viahqx')
516 #hexbin(fname, fname+'.unpacked')
519 if __name__ == '__main__':